Physical Address
23,24,25 & 26, 2nd Floor, Software Technology Park India, Opp: Garware Stadium,MIDC, Chikalthana, Aurangabad, Maharashtra – 431001 India
The rapid expansion of artificial intelligence is raising fresh environmental concerns, and not just because of its heavy electricity and water use. A new study suggests that large AI data centres are also heating the land around them, creating localised “heat islands” that could affect more than 340 million people worldwide. The research, led by scientists at the University of Cambridge, found that once a major AI data centre begins operations, surrounding land surface temperatures rise by an average of 2°C, with some areas recording increases of up to 9°C. The findings point to a growing but largely overlooked climate cost of the AI boom, especially in already heat-stressed regions.
AI infrastructure is creating a new kind of urban heat
The study analysed more than 20 years of satellite-based land surface temperature data and compared it with the locations of large AI data centres around the world. Researchers found that these facilities, many of which cover more than 93,000 square metres and house thousands of servers, consistently raised temperatures in surrounding areas after beginning operations. On average, the rise was 2°C, but in extreme cases it reached 9°C, a difference large enough to transform local thermal conditions in ways that communities can feel.
The warming effect was not limited to the immediate edge of a facility. According to the research, elevated temperatures extended as far as 10 kilometres from some data centres. The study describes this as a “data heat island effect,” similar in principle to the urban heat island effect seen in cities where concrete, asphalt and infrastructure trap heat. In this case, however, the heat source is tied to the growing concentration of computing infrastructure required to run tools such as large language models and other AI systems.
From Spain and Mexico to India, the impact is already visible
The pattern showed up across multiple regions. The study points to places such as Bajío in Mexico and Aragón in Spain, both major data centre hubs, where unexplained temperature increases of around 2°C were recorded in areas around AI infrastructure but not in neighbouring zones. That consistency across geographies suggests the effect is not isolated to one climate or one kind of built environment.
For India, the findings are particularly relevant because the country is rapidly scaling up its digital infrastructure even as many of its cities are already facing dangerous summer heat. The environmental footprint of this growth is likely to increase sharply. According to testimony recorded in the 27th Report of the Standing Committee on Communications and Information Technology, tabled in Parliament on Monday, India’s data centres currently consume around 1,020 megawatts of power. That demand is expected to double within two years, cross 2,000 megawatts, and could rise to 4 to 5 gigawatts within four to five years, according to the Ministry of Electronics and Information Technology.
Why this matters in a warming world
Heat is not just a discomfort issue. In a warming climate, even small increases in local temperatures can worsen health risks, raise cooling demand and make already vulnerable neighbourhoods harder to live in. In India, where heatwaves are becoming more frequent and more dangerous, any additional local warming can add pressure to public health systems, electricity demand and urban planning.
The researchers argue that the environmental costs of AI can no longer be discussed only in terms of carbon emissions, electricity demand or water use. The physical heat released by large-scale data infrastructure is also becoming part of the story. As governments and companies race to build more AI capacity, the study suggests that the placement, design and cooling of data centres will need much closer scrutiny, especially in regions that are already on the front line of extreme heat.
References:
https://arxiv.org/pdf/2603.20897
Banner image: Photo by Taylor Vick on Unsplash
Sections of this article may have been developed with the assistance of artificial intelligence tools to support research, drafting and language refinement. All information has been reviewed, edited and verified by the author/editor to ensure accuracy, context and editorial integrity. The responsibility for the final content, interpretations and conclusions rests solely with the publisher.
